routelint

A command-line tool for finding broken links, incorrect redirects and indexing conflicts on a live or preview website. It discovers URLs, checks what your server returns and gives you findings tied to specific pages, with the evidence behind them.

DownloadVersion 0.4.0

routelint

About routelint

Moving a page is easy enough. Remembering every link to it, sending the old URL to the right replacement and keeping the sitemap and canonical tags in agreement is where things get less fun. RouteLint helps check those details after a deployment or during a migration.

Start with a site URL, give it a list of routes, or let it discover pages through sitemaps and internal links. It looks for broken destinations, redirect chains, missing metadata, empty server-rendered pages and conflicting indexing instructions. It also checks how pages connect, including dead ends and pages buried several clicks deep.

For migrations, you can tell it exactly where an old URL should go, which redirect status to expect and how many hops are acceptable. It then checks that the redirect reaches the intended destination and that the destination actually works. “It redirects somewhere” is a fairly low bar.

Normal checks use the HTML and headers returned by the server. If you want to see what changes after JavaScript runs, you can enable the optional Playwright comparison. That helps spot content or SEO tags that are missing from the original response but appear in the browser.

You can save a report before making changes and compare it with the next run, or show only new and worsening findings. Useful when you already know about a backlog and want to see what this particular release has added to it.

It’s written in TypeScript and runs locally or in CI. Reports are available in the terminal, as a shareable HTML file, or in formats for automated checks. You control the crawl limits and which findings should fail a run.

Release history

Version 0.4.021 Sep 2026

Added

Opt-in response-time budgets. audit.maxResponseMs reports slow-route findings and audit.maxRedirectHopMs reports slow-redirect findings, each carrying the observed duration, the configured budget, and the route or hop involved. Both are disabled until a threshold is configured and can be narrowed or widened per path scope.
Redirect pattern contracts. A source may use * for one path segment and a trailing ** for the remaining path; each placeholder in to is replaced in order, so /blog/old/* → /blog/new/* is verified against matching crawl URLs and explicit samples. Samples enter the crawl subject to its filters and limits; unfetched samples are unchecked. redirect-pattern-unmatched records a pattern without matching sources.
Query strings in redirect contracts, and destinations that leave the audited origin. Cross-origin targets are verified through the source's own redirect chain instead of being added to the same-origin crawl.
Markdown, CSV, and JUnit report formats. --format markdown writes a GitHub step summary, --format csv writes a spreadsheet-friendly table of findings, and --format junit writes JUnit XML for CI test reporters. Markdown and XML output neutralize report values instead of embedding them as markup.
Crawl pacing. Crawl-delay and Request-rate records in robots.txt now space raw page request starts, including redirect hops and agent comparisons, with the declared delay capped at 10 seconds. limits.delayMs and --delay <duration> set uncapped explicit minimum spacing, limits.honorCrawlDelay and --ignore-crawl-delay control whether the declared value is honored, and a crawl-policy finding records the pacing that was applied or warns when a declared delay had to be capped. Pacing waits do not consume capture timeout or response-time budgets.

Changed

JSON report schema is now version 4. The report reader and diff command remain compatible with schemas 1, 2, and 3.
Redirect contract checks are counted per matched source, so redirectContracts adds patterns, patternMatches, and unmatchedPatterns, and each check records the pattern it came from.
Report configuration snapshots record delayMs and honorCrawlDelay.
The starter and example configurations document the new budgets, pacing options, and pattern contracts.

Fixed

Preserve exact query ordering and encoding throughout redirect candidate collection; retain trailing slashes in ** captures and merge duplicate pattern samples.
Redact sensitive values in redirect samples and pattern fields, and preserve per-agent robots delay evidence in serialized reports.
Neutralize Markdown formatting and spreadsheet formulas in report values, remove invalid XML characters from JUnit output, and show unmatched patterns even when no concrete checks ran.
Apply redirect response-time budgets to each hop's own path scope, compare pacing settings in changed-only reports, and accept --delay 0.

External download

Download 0.4.0

Version 0.3.18 Sep 2026

Fixed

Ignore links, metadata, headings, and base URLs inside HTML <template> elements, including nested templates. Unused template content no longer creates phantom crawl routes or false SEO findings.

Changed

Rewrite the README and package description with clearer explanations and practical examples.
Update the publishing guide to release through a pull request to protected main, then publish to npm interactively from the merged commit.

0.3.0 - 2026-08-27

Added

Exact same-origin redirect contracts with first-hop status, final destination, hop-limit, target-status, and target-indexability checks.
Automatic live validation for concrete, unconditional, query-free, same-origin redirects found in supported Next.js build manifests.
Structured redirect-contract results in JSON, a dedicated HTML table, terminal outcome totals, SARIF findings, and public TypeScript APIs.

Changed

JSON report schema is now version 3. The report reader and diff command remain compatible with schemas 1 and 2.
Correctly declared redirect sources no longer produce the generic redirected-route or ordinary page-response findings; sitemap and internal-link findings still apply.
Configured redirect contracts participate in changed-only baseline comparability.

External download

Download 0.3.1

Version 0.2.024 Aug 2026

Added

Optional Playwright Chromium pass that compares raw SSR evidence with the rendered DOM while injecting configured headers only into audited-origin requests.
Body-text evidence with word/character counts, SHA-256, and SimHash; reports do not retain page text.
Empty SSR shell, client-only content, rendered-only SEO signal, soft-404, possible-soft-404, and duplicate SSR body checks.
Explicit URL-list files and stdin input with source-line diagnostics, same-origin enforcement, size limits, and stable de-duplication.
Global and path-scoped requirement/severity configuration, including disabling finding codes.
--changed-only output for new or worsened findings against a schema-validated baseline.
Dedicated agent-response comparison in HTML reports and clearer collection scope in terminal/HTML output.
Representative frozen Next.js 16 App Router manifests and Astro SSR output fixtures.

Changed

JSON report schema is now version 2. The report reader and diff command remain compatible with schema 1.
Browser rendering is an optional peer dependency; default installs remain raw-HTTP only and do not download Chromium.
Runtime and development dependencies use the updated versions supplied in the Dependabot-clean 0.1.0 baseline.
CI remains verification-only. npm publishing is still an interactive local operation protected by account and package 2FA.

External download

Download 0.2.0

Version 0.1.022 Aug 2026

Bounded same-origin crawler with explicit redirects, response limits, bot/browser delivery comparisons, and credential redaction.
Recursive sitemap, robots.txt, internal-link, and Next.js build-manifest route discovery.
Server-rendered metadata, canonical, robots, H1, language, link, and hreflang capture.
Cross-route sitemap, indexability, canonical, hreflang, link-graph, depth, orphan, duplicate, and route-variant checks.
Terminal, JSON, SARIF, and self-contained HTML reports.
Schema-validated baseline report diffs.
Fail-closed robots handling for unavailable policies and API/route-handler filtering in Next.js discovery.
Cloudflare email-protection pseudo-links excluded from route discovery.
Strict TypeScript API, CLI, cross-platform package smoke tests, and GitHub Actions CI without npm publishing credentials.

External download

Download 0.1.0

routelint

About routelint

Release history

More software